Timeline for How to Generate Band Limited Gaussian White Noise in MATLAB?
Current License: CC BY-SA 4.0
12 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 7, 2022 at 19:17 | comment | added | Ian Bland | And your solution was clearly under-appreciated @Zibri ;) My current application was on Windows but as also a microcontroller programmer I agree that people just don't seem to appreciate light, compact, efficient code any more. I do at least! | |
| Aug 7, 2022 at 10:33 | comment | added | Zibri | @IanBland you're welcome ;) | |
| Jul 28, 2022 at 18:16 | comment | added | Ian Bland | Anyone else who finds this; I've just used Zibri's method to generate variable Q noise for a software synthesizer project and it works very well and is lighter weight than trying to do it with filters. It was exactly what I was looking for. | |
| Sep 28, 2019 at 6:43 | comment | added | Zibri | I feel like nobody here knows how to write good code. Saying that a biquad filter is as fast as this is just ignorant. When needing realtime sound generation expecially on microcontrollers you need the shortest and fastest code. Today it seems nobody can do anything without using some huge library or some huge all purposes routine. Bye bye, and thanks for nothing. | |
| Sep 24, 2019 at 13:43 | comment | added | user28715 | since your technique so easy to modify, why don’t you write your own sine wave routine instead of linking to one and claiming that your code is less complex and compare that to a biquad. while you are add it, how about a fixed point version. you might also think about how your technique scales as you increase the number of samples | |
| Sep 24, 2019 at 12:32 | comment | added | user28715 | Actually the inband ripple is a design parameter. To say they peak at the center frequency is also false | |
| Sep 24, 2019 at 10:29 | comment | added | Zibri | Also filters don't give a flat output but they peak at the center frequency... anyway that's exactly what I needed in my project... to alter the routine and make it flat is very easy. | |
| Sep 23, 2019 at 11:37 | comment | added | user28715 | your noise isn’t preferable. the spectrum isn’t flat across the band. calling a transcendental function on the fly isn’t faster than filtering. your arguments are unsubstantiated | |
| Sep 23, 2019 at 6:21 | comment | added | Zibri | @StanleyPawlukiewicz In electronics you are righe because a noise generator and a filter are two very simple "objects". In programming, a filter is complex in number of instructions way more than just generating the needed data. So when programming is preferable my approach if you have time critical applications. If you downvoted I suggest you to think again about it. | |
| Sep 23, 2019 at 6:19 | history | edited | Zibri | CC BY-SA 4.0 |
added 68 characters in body
|
| Sep 22, 2019 at 12:59 | comment | added | user28715 | The STFT isn’t flat across the band. You have not shown why there is any advantage to your technique. Btw most paint is stocked as a grayish white and then mixed with pigment. Orange paint isn’t made by just using orange ingredients. There isn’t any extra work in generating white noise. | |
| Sep 22, 2019 at 9:36 | history | answered | Zibri | CC BY-SA 4.0 |